Pathway 3: "I Want to Build AI Agents" (Software Engineer)

Target audience: Software engineers who want to build autonomous AI agent systems

Goal: Build autonomous agents that use tools, maintain state, coordinate with other agents, and interact with users through modern protocols (MCP, AG-UI).

Chapter Guide

• Start Ch 00: ML and PyTorch Foundations (review if new to ML) review ML basics if new to the field
• Start Ch 01: NLP and Text Representation (review if new to ML) review NLP basics if new to the field
• Focus Ch 10: Working with LLM APIs the API layer your agents call
• Focus Ch 11: Prompt Engineering craft system prompts that guide agent behavior
• Skim Ch 08: Reasoning Models and Test-Time Compute reasoning models power complex agent planning
• Focus Ch 19: Embeddings and Vector Databases agents need retrieval for knowledge grounding
• Focus Ch 20: RAG give your agents access to real documents
• Focus Ch 21: Conversational AI handle multi-turn conversations with state
• Focus Ch 22: AI Agent Foundations core theory: planning, memory, and tool use
• Focus Ch 23: Tool Use and Protocols (MCP, function calling) implement MCP, function calling, and AG-UI
• Focus Ch 24: Multi-Agent Systems coordinate multiple agents on complex tasks
• Focus Ch 25: Specialized Agents (code, research, data) build code agents, research agents, data agents
• Focus Ch 26: Agent Safety and Production Infrastructure safety guardrails and production patterns
• Skim Ch 28: LLM Applications see how agents fit into real applications
• Focus Ch 29: Evaluation and Experiment Design evaluate agent reliability and correctness
• Skim Ch 30: Observability and Monitoring trace and debug agent execution flows
• Skim Ch 31: Production Engineering deploy and scale agent systems
• Skim Ch 32: Safety, Ethics and Regulation compliance and safety for autonomous systems
• Skim Ch 34: Emerging Architectures next-generation architectures for agent capabilities
• Optional Ch 35: AI and Society societal context for responsible agent design